case the failure is usually more or less abrupt. The former type of failure is more common in fireclay, the latter with silica, chromite and magnesite refractories.
The test as usually carried out is very simple and consists in applying pressure upon a standard size firebrick, placed on end within a suitable furnace, and raising the temperature at the prescribed rate. The maximum temperature is maintained con-
stant for some time, usually 1J^ hr. The furnace recommended for this purpose is shown in Fig. 5, in which provision is made for the simultaneous testing of two specimens. The expansion and contraction may be followed by attaching a pantograph recorder to the beam but the specimen is rated according to the contraction it has uffered expressed in per cent of the original length or whether it has failed by collapse. The maximum temperature to be employed depends upon the kind of refactory to be tested. It is usually taken to be 1,350°C. for grade one firebrick, 1,400°C. for silica